We request renewal of support for this program in biometric research in cancer epidemiology. The aim of the program is to develop new technology or to extend previously developed technologies for scientific data-utilization. This includes computerized biostatistical and mathematical models capable of handling the complexities of epidemiological study situations. It also includes the application of the natural language technology developed in modern linguistics to improve each of the 9 steps in the process of data-utilization. This new technology is designed to work on real (as opposed to hypothetical or ideal) data and to product practical results in primary prevention that would help immediately to cut cancer death rates in future years. The technology would be initially applied to the following primary research targets: (1) further studies of the health effects of low-level ionizing radiation, (2) follow-up studies of the iatrogenic diseases that may have been produced by the surgical adjuvant treatments for breast cancer, (3) applications of natural language computing systems to recoding of Tri State Survey date, developing of computerized archives, and planning a largely automated procedure for rigorously assessing the validity of a scientific argument. After a two-year rebuilding period, further lines of research will be considered.